Abstract:
|
Degradation tests have been widely used to assess the lifetime information of highly reliable products. In this paper, we consider the degradation tests based on Wiener processes. In practice, there may exist unit-to-unit variation among products. Therefore, it is not appropriate to use the same model to fit all products. We aim on the Bayesian approach with three different models. Empirical Bayesian methods are used to determine the prior parameters using all the observed data. It not only provides a compromise in the model uncertainty with the advantage of "borrowing the strength" among the data, but also solves the situation when the prior information is vague. The ultimate goal is to extrapolate the lifetime distribution from the degradation model. Simulation study shows that the empirical Bayesian method can reduce the risk in model fitting and yield more robust reliability inference.
|